CN103995823A - Information recommending method based on social network - Google Patents

Abstract

The invention discloses a social network-based information recommendation method, the steps of which are as follows: step 1: calculate the trust degree and similarity between users, and use weighted values to construct a user relationship matrix; step 2: use a community discovery algorithm to identify users Carry out clustering to form the user's nearest neighbor set; Step 3: predict the score and generate a recommendation list. The present invention can achieve the following beneficial effects: (1) Solve the problem of cold start. The present invention introduces the degree of trust. If there are not enough neighbors based on the common scoring items when recommending, trusted friends can be used as the starting point for prediction, which can alleviate the cold start problem and improve user coverage; (2) Improve real-time performance. In the present invention, the community discovery algorithm commonly used in social network analysis is used to divide the user network into communities, that is, the same user interests are clustered, so that the time for searching for user neighbor sets is greatly shortened, and the real-time response of information recommendation is improved.

Description

A method of information recommendation based on social network

technical field

The present invention relates to the technical field of network information, in particular to an information recommendation method based on a social network.

Background technique

The rapid development of the Internet and the ever-increasing information resources have led to a sharp increase in the information index. The information service field is facing the problem of "rich information resources, but difficult to obtain useful information", which brings a great information burden to people. On the one hand, there is an "information overload" phenomenon (information overload) caused by a large number of data resources on the network; on the other hand, users cannot obtain the information resources they need. Recommendation systems (recommendation systems), as an important method of "information push" mode of service, are the main means to solve the problem of information overload. For information that is difficult to obtain, recommend more valuable information resources for users according to their behavior characteristics in different environments.

Along with the development of the Internet is the rapid expansion of social networks. Social networks connect people with the same hobbies or even strangers through the Internet to form groups with certain characteristics. Social network is a platform that can communicate with each other and participate in interaction. It has expanded from research departments, schools, governments, and business application platforms to a tool for human social communication. Because the Internet is virtual, people do everything possible to conceal their real identities on the Internet, which not only brings a lot of false information, but also makes the level of trust between people lower and lower, and communication becomes more difficult. The social network uses real information registration, which enhances the identity authenticity and behavioral credibility of network users, greatly guarantees the reliability, regionality and real-time nature of information security and user interaction in the system, and makes people feel more at ease and relaxed Communicating with other people, but also brings a new user experience. Through social networks, they will actively announce their characteristics and preferences, actively provide and annotate various resources (such as pictures, videos) or share their knowledge. For example, users can share books through Douban, conduct social networking and share photos through Facebook, send Weibo through Twitter, post photos through Flickr, and upload videos through YouTube. The increasingly popular social network is quietly changing people's lifestyle and value orientation.

Currently, commonly used recommended methods include the following:

1) The recommendation method based on association rules, which generates association rules based on user transaction data, and puts forward suggestions based on the user's current purchase behavior. Shopping cart analysis is the most typical application of association rules. The recommendation method based on association rules has strong versatility and can be applied in many fields, but it is difficult to extract association rules and consumes a lot of time. With the increasing number of association rules, the system becomes difficult to manage.

2) Content-based recommendation method, which mainly focuses on the content analysis of information resource items and the construction of user preference model. The recommendation function is realized by comparing the similarity between resources and user preferences. Although the content-based recommendation technology has intuitive results, simple calculation, fast response time, and good explainability, it can solve the problems of cold start and data sparseness. However, there are still some limitations: the content that can be analyzed is limited, it is only information that can be represented by a series of feature sets, and it cannot effectively process multimedia information such as sound, pictures, and video; users can receive and Liked to recommend similar items in the past, but could not discover new and interesting products for users, and the recommended content was single; unable to deal with quality, style or opinion.

3) Collaborative filtering recommendation method, which is currently the most successful technology in recommending information systems. The basic idea of collaborative filtering is to use the similarity between users or items to make recommendations or predictions. This method finds a group of users with the same preferences , and then analyze the common preferences of users to make recommendations to target users. The advantage of the collaborative filtering algorithm is that it does not pay attention to the content of the project itself, but mainly recommends resources based on the similarity of users or projects. The system only needs to obtain enough project evaluations to make project recommendations reliably. However, the disadvantages of collaborative filtering algorithm are also very obvious, namely "cold start" problem, data sparse problem, scalability problem and so on.

The development of social network provides a good channel for personalized recommendation. The present invention organically combines the trust measurement between users, social network mode and personalized recommendation technology, and proposes an information recommendation method based on social network to construct a high-efficiency , High-precision recommendation system.

Contents of the invention

In order to solve the above-mentioned technical problems, the present invention provides a social network-based information recommendation method, and the technical solution adopted is as follows:

A social network-based information recommendation method, the steps of which are as follows:

Step 1: Calculate the trust and similarity between users, and use the weighted value to construct the user relationship matrix;

Step 2: Use the community discovery algorithm to cluster users to form a user nearest neighbor set;

Step 3: Predict ratings and generate recommendation lists.

Construct user-rating matrix R _m×n :

m represents the number of users, n represents the number of items, r _ij represents the rating of user i on item j.

On the basis of the filled user-score matrix R _m×n , use Pearson correlation to calculate the similarity between users, and construct a user-user similarity matrix S:

s _uv represents the similarity between user u and user v, and s _uv ∈ [0,1].

Calculate the trust degree, and then calculate the weighted value combining similarity and trust:

h(i,j)=θ×trust(i,j)+(1-θ)×sim(i,j);

θ represents the weighting parameter, trust(i,j) represents the direct trust degree of users i and j, sim(i,j) represents the similarity, h(i,j) represents the weighted value of the two.

Use the weighted value to construct the user relationship matrix H:

The direct trust degree is the combination of the interactive trust degree, the proportion of common friends between users, and the user's evaluation ability. Its calculation formula is:

$\mathrm{<span\; class="notranslate">\; trust</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; fri</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; fre</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; comm</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

trust(i,j) represents the direct trust between users i and j, fri(i,j) represents the proportion of common friends between users, fre represents the evaluation ability of users, and comm(i,j) represents the relationship between users interaction trust.

The calculation formula of mutual trust degree is as follows:

$\mathrm{<span\; class="notranslate">\; comm</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; out</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; ;</span>$

Among them, w _i,j represents the number of messages sent by user i to user j, and Σw _i(out) represents the total number of messages sent by user i to surrounding users;

The calculation formulas for the proportion of common friends between users and the evaluation ability of users are as follows:

$\mathrm{<span\; class="notranslate">\; fri</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&cap;</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; ;</span>$

$\mathrm{<span\; class="notranslate">\; fre</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; ;</span>$

Among them, fri(i,j) represents the proportion of common friends between users, fre represents the evaluation ability of users, n _i and n _j represent the number of friends of users i and j respectively, and n _i ∩ n _j represents their common The number of friends; k _i represents the number of times the recommended user evaluates the product category i, and Σk _i represents the number of times the user evaluates all types of products.

The process of community discovery is as follows:

Step1: Calculate the degree of each user in the network (the number of edges associated with the vertex), and select the user i with the highest degree as the initial community C _i , and initialize the modularity Q=0;

Step2: find out all the users connected with the community C _i , and put them into the adjacent user set N;

Step3: Calculate the contribution q of each user j in the user set N to the community C _i , and add the user with the greatest contribution to the community to the community C _i ;

Step4: Calculate the modularity Q' of the community C _i . If Q'>Q, add user j to community C _i successfully, mark user j, and update modularity Q=Q' at the same time, return to Step 2 to continue execution; otherwise, go to Step 5;

Step5: The modularity Q has reached the maximum value, that is, the current community C _i has reached the optimal result of division;

Step6: If there are no unmarked users and all communities in the network have been detected, the process ends; otherwise, select the user with the highest degree from the unmarked users as the new initial community C _i , and return to step2 to continue .

The user's contribution to the community q

$\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; in</span>}}}{{\mathrm{<span\; class="notranslate">\; l</span>}}_{\mathrm{<span\; class="notranslate">\; in</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; out</span>}}}<span\; class="notranslate">\; ;</span>$

L _in : represents the number of connected edges within the community in the unweighted network; represents the sum of the weights of all edges within the community in the authorized network.

L _out : In the unweighted network, it represents the number of external edges connected to the community; in the authorized network, it represents the sum of the weights of all the external edges connected to the community.

The greater the contribution q of the user to the community, the closer the connection between the user and the community.

Definition 2: Modularity Q

Modularity Q is an important indicator to measure whether the current community has reached the best level in community division. The larger the modularity Q is, the better the community division effect is, which means that the connection between the community and the community is less, so as to realize the "internal" High cohesion" and "low coupling" outside the module.

If it is an unweighted network, the expression of modularity Q is as follows:

$\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; m</span>}}\underset{\mathrm{<span\; class="notranslate">\; ij</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Among them, m is the total number of edges _of the network; k _i and k _j respectively represent the number of connection edges with user i and user _j ; =1, when user i and user j are not connected, A _ij =0; δ(C ⁱ ,C _j ) is a kronecker function, if user i and user j belong to the same community, then δ(C _i ,C _j )= 1. If they are not in the same community, then δ(C _i , C _j )=0.

In a weighted network, the modularity Q can be defined as follows:

$\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; W</span>}}\underset{\mathrm{<span\; class="notranslate">\; ij</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; W</span>}}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Among them, W represents the sum of the weights of all edges in the network, and w _i and w _j represent the sum of the weights of the edges connected to user i and user j respectively.

According to the score information of the target user's n nearest neighbors on the candidate items, predict the target user's score on the candidate items, and select the top few items with the highest predicted scores, and actively push them to the target user as the recommendation result, that is, generate top-N information Resource recommendation.

${\mathrm{<span\; class="notranslate">\; P</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; +</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; trust</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}<span\; class="notranslate">\; )</span>}{\underset{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; trust</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ;</span>$

Among them, P _u,i represents the predicted score of user u on item i, and Represents the average ratings of user u and neighbor user v on items respectively; R _v,i represents the rating of user v on item i, trust(u,v) represents the trust degree of user u to neighbor user v, N represents the neighbors of user u user candidate set.

The direct trust degree trust(u,v) can also be replaced by the transfer trust degree trust _L(A,B) , and its calculation method is as follows:

trust _L(A,B) ＝trust(A,X ₁ )×trust(X ₁ ,X ₂ )×…×trust(X _n ,B);

Among them, _Xi represents the user between user A and user B on the path L, and L(A, B) represents the existing trust path between user A and user B. If there are multiple trust paths between user A and user B in the trust network trust paths L(L ₁ ,L ₂ ,…,L _n ),(n≥2), select the shortest path in path L, if there are k shortest paths, the calculation formula is as follows:

${\mathrm{<span\; class="notranslate">\; trust</span>}}_{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; trust</span>}}_{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}}{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span><span\; class="notranslate">\; ;</span>$

Then continue to calculate according to the order of path L first short and then long, if the calculation result of a certain path j Then the path j is counted as one of the shortest paths, and trust _L(A,B) is recalculated, and the final indirect trust result is obtained after all paths are calculated.

The information recommendation method based on social network proposed by the present invention can achieve the following beneficial effects:

(1) Solve the cold start problem. The cold start problem means that when a new user joins the recommendation system, since there is no record of the new user's historical behavior or rating, it is impossible to make effective information recommendations for the new user. This method introduces the degree of trust. If there are not enough neighbors based on the common scoring items when making recommendations, trusted friends can be used as the starting point for prediction, which can alleviate the cold start problem and improve user coverage.

(2) Improve real-time performance. In this method, the community discovery algorithm commonly used in social network analysis is used to divide the user network into communities, that is, the same user interests are clustered, which greatly shortens the time when searching for user neighbor sets and improves the real-time response of information recommendation.

Description of drawings

Figure 1 is a flowchart of a recommendation method based on social networks.

Figure 2 Interaction diagram between users.

Fig. 3 Multi-path transitive trust calculation diagram.

Figure 4. Flowchart of community discovery.

Detailed ways

The invention proposes an information recommendation method based on a social network. First, calculate the trust and similarity between users, and use the weighted value to construct a user relationship matrix; second, use a community discovery algorithm to cluster users to form user nearest neighbor sets; finally, predict ratings and generate recommendation lists.

As shown in Figure 1, collect relevant information such as users, items, ratings, and social networks, calculate the trust and similarity between users based on these information, and use the weighted value to construct the user relationship matrix; secondly, use the community discovery algorithm to Users are divided to form user nearest neighbor sets; finally, prediction scores are generated and recommendation lists are generated.

1. Direct trust

Immediate trust is a quantification of how much one user trusts another user. To introduce trust into a personalized recommendation system, the first thing to be solved is the quantification of trust relationship, making it a data that can be calculated or performed related operations. Combining multiple properties of trust and related characteristics of actual social networks, this paper considers the calculation of direct trust degree from the following aspects.

1) Interaction trust

The degree of interaction refers to the degree of interaction between users in the social network relationship, such as the number of interactions in a period of time such as user calls, text messages, emails, microblog followers, and SNS messages. The interaction graph between users is shown in Figure 2. Each node in the graph represents the users in the social network, and the edges between users represent the interactive sender pointing to the receiver with directed arrows, and the weight values on the edges represent various interaction types. sum of information. Through the interaction graph between users, we can define the interactive trust between users as:

$\mathrm{<span\; class="notranslate">\; comm</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; out</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; ;</span>$

Among them, w _i,j represents the number of messages sent by user i to user j, and Σw _i(out) represents the total number of messages sent by user i to surrounding users.

2) Recommendation trustworthiness

Recommendation trust includes the proportion of common friends between users and the evaluation ability of users. Users always trust those users who have purchased or used the product more. If their background information such as their own profession and interests has a certain relationship with the product, their own background makes them have a certain evaluation of this type of product. ability, which can also be a reference factor for the reliability of its recommendation of the product. At the same time, the number of common friends between users can also reflect the relationship between the two. fri(i,j) represents the proportion of common friends between users, and fre represents the user's evaluation ability. The calculation is as follows:

$\mathrm{<span\; class="notranslate">\; fri</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&cap;</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; ;</span>$

$\mathrm{<span\; class="notranslate">\; fre</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; ;</span>$

Among them, n _i and n _j represent the number of friends of users i and j respectively, and n _i ∩ n _j represents the number of their common friends; _{k i} represents the number of times recommended users evaluate products of category i; of evaluations.

The direct trust degree trust(i,j) can be obtained by combining the above two points of trust:

$\mathrm{<span\; class="notranslate">\; trust</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; fri</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; fre</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; comm</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

2. Transfer trust calculation

In the transfer calculation of the trust degree, the principle that direct trust takes precedence over indirect trust should be followed, and L(A,B) is used to represent the existing trust path between user A and user B. If there is a unique trust path L(A,X ₁ ,X ₂ ,…,X _n ,B) between user A and user B in the trust network, then the transfer trust degree between user A and B is on the trust path L The product of all direct trust degrees of , the expression is:

trust _L(A,B) ＝trust(A,X ₁ )×trust(X ₁ ,X ₂ )×…×trust(X _n ,B);

Among them, _Xi represents the user between users A and B on the path L.

If there are multiple trust paths L(L ₁ ,L ₂ ,…,L _n ),(n≥2) between user A and user B in the trust network, then select the shortest path in path L, if there are k The shortest path, the calculation formula is as follows:

${\mathrm{<span\; class="notranslate">\; trust</span>}}_{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; trust</span>}}_{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}}{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; min</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span><span\; class="notranslate">\; ;</span>$

Then continue to calculate according to the order of path L first short and then long, if the calculation result of a certain path j Then path j is counted as one of the shortest paths, and trust _L(A,B) is recalculated. When all paths (if the link is too long, the calculation can be ignored), the final indirect trust result is obtained. This calculation method effectively avoids misjudgment caused by too low trust in the shortest trust path.

As shown in Figure 3, there are multiple trust paths from user A to user B. First, the trust _L( A,B) of the shortest path L(A,X ₃ ,B) and L(A,X ₄ ,B) should be calculated value, But the transitive trust degree of another path L(A,X ₁ ,X ₂ ,B) ${\mathrm{trust}}_{L(A,x_1,x_2,B)}=0.8\&times;0.8\&times;0.9=0.576>{\mathrm{trust}}_{L\left(\text{A,B}\right)}=0.48,$ Therefore, the path L(A,X ₁ ,X ₂ ,B) should also be added to the shortest path set, and the final transfer trust degree is

3. Community Discovery

In the real network, there are often some users who are very closely connected with other users (that is, the point with a large user degree in the network, which we call "central user"). We regard the central user as the starting point of the initial community, and then continue to add adjacent users who contribute the most to the community to the community (if there are multiple users who contribute greatly to the community, all these users will be added to the community Middle), a stable community is formed when the global contribution reaches the maximum.

In order to understand the clustering process more clearly, the community contribution and modularity are defined as follows:

Definition 1: User's contribution to the community q

$\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; in</span>}}}{{\mathrm{<span\; class="notranslate">\; l</span>}}_{\mathrm{<span\; class="notranslate">\; in</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; out</span>}}}<span\; class="notranslate">\; ;</span>$

L _in : represents the number of connected edges within the community in the unweighted network; represents the sum of the weights of all edges within the community in the authorized network.

L _out : In the unweighted network, it represents the number of external edges connected to the community; in the authorized network, it represents the sum of the weights of all the external edges connected to the community.

An unweighted network means that there is no weight value on the edge, and a weighted network means that there is a weight value on the edge.

The greater the contribution q of the user to the community, the closer the connection between the user and the community.

Definition 2: Modularity Q

Modularity Q is an important indicator to measure whether the current community has reached the best level in the community division. The larger the modularity Q, the better the community division effect, which means that the connection between the community and the community is less, so as to realize the "internal" of the module. High cohesion" and "low coupling" outside the module.

If it is an unweighted network, the expression of modularity Q is as follows:

$\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; m</span>}}\underset{\mathrm{<span\; class="notranslate">\; ij</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Among them, m is the total number of edges _{of the} network; k _i and k _j respectively represent the number of connection edges with user i and user j; =1, when user i and user j are not connected, A _ij =0; δ(C _i ,C _j ) is a kronecker function, if user i and user j belong to the same community, then δ(C _i ,C _j )= 1. If they are not in the same community, then δ(C _i , C _j )=0.

In a weighted network, the modularity Q can be defined as follows:

$\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; W</span>}}\underset{\mathrm{<span\; class="notranslate">\; ij</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; W</span>}}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Among them, W represents the sum of the weights of all edges in the network, and w _i and w _j represent the sum of the weights of the edges connected to user i and user j respectively.

The community discovery algorithm is shown in Figure 4:

Step1: Calculate the degree of each user in the network (the number of edges associated with the vertex), and select the user i with the highest degree as the initial community C _i , and initialize the modularity Q=0;

Step2: find out all the users connected with the community C _i , and put them into the adjacent user set N;

Step3: Calculate the contribution q of each user j in the user set N to the community C _i , and add the user with the greatest contribution to the community to the community C _i ;

Step4: Calculate the modularity Q' of the community C _i . If Q'>Q, add user j to community C _i successfully, mark user j, and update modularity Q=Q' at the same time, return to Step 2 to continue execution; otherwise, go to Step 5;

Step5: The modularity Q has reached the maximum value, that is, the current community C _i has reached the optimal result of division;

Step6: If there are no unmarked users and all communities in the network have been detected, the process ends; otherwise, select the user with the highest degree from the unmarked users as the new initial community C _i , and return to step2 to continue .

4. Recommended method implementation

1) Build a user relationship matrix

First, obtain user information, item information, and rating information, and use a matrix to represent the relationship between users and items to form a user-rating matrix R _m×n , where item information refers to products, movies, etc. recommended to users thing. Among them, m represents the number of users, n represents the number of items, r _ij represents the rating of user i on item j. To reduce the sparsity of the user-rating matrix, unrated r _ij can be set to 0 or the mean of all user i rated or item j rated.

Secondly, on the basis of the filled user-rating matrix R _m×n , use Pearson correlation to calculate similarity. Pearson correlation is a method to calculate similarity. As I _ij , the similarity sim(i,j) measured by Pearson correlation is:

$\mathrm{<span\; class="notranslate">\; sim</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\underset{{\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; ic</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; jc</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; )</span>}{\sqrt{\underset{{\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; ic</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&times;</span>\underset{{\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; jc</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}<span\; class="notranslate">\; ;</span>$

Among them, R _ic and R _jc represent the ratings of two users on item c; and Represent the average ratings of user i and user j on I _ij respectively.

On the basis of the above, a user-user similarity matrix S is constructed. Among them, s _uv represents the similarity between user u and user v, and s _uv ∈ [0,1].

User-user similarity relationship matrix S:

Calculate the trust degree, and then calculate the weighted value combining similarity and trust:

h(i,j)=θ×trust(i,j)+(1-θ)×sim(i,j);

θ represents the weighting parameter, trust(i,j) represents the trust degree of users i and j, sim(i,j) represents the similarity, and h(i,j) represents the weighted value of the two.

Construct the user relationship matrix H:

2) Clustering users with the same interest to form a neighbor set

According to the user relationship matrix, users are clustered using the community discovery algorithm described above, and users are divided into several communities {C ₁ ,C ₂ ,…,C _n }, and users in each community have the same hobbies , forming a neighbor set. If a new user joins, the user can also be reclassified by updating the network.

3) Predict scoring and generate recommendations

According to the score information of the target user's n nearest neighbors on the candidate items, predict the target user's score on the candidate items, and select the top few items with the highest predicted scores, and actively push them to the target user as the recommendation result, that is, generate top-N information Resource recommendation.

${\mathrm{<span\; class="notranslate">\; P</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; +</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; trust</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}<span\; class="notranslate">\; )</span>}{\underset{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; trust</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ;</span>$

in, and Represents the average ratings of user u and neighbor user v on the item respectively; trust(u, v) represents the trust degree of user u to neighbor user v.

The direct trust degree trust(u,v) can also be replaced by the transfer trust degree trust _L(A,B) , and its calculation method is as shown above.

Claims (8)

1. An information recommendation method based on a social network comprises the following steps:

step 1: calculating the trust and similarity between users, and constructing a user relationship matrix by using a weighted value;

step 2: clustering the users by using a community discovery algorithm to form a user nearest neighbor set;

and step 3: the prediction scores and generates a recommendation list.

2. The information recommendation method based on social network as claimed in claim 1, wherein the user relationship matrix is constructed in step1 by the following method:

constructing a user-score matrix R_m×n：

m represents the number of users, n represents the number of items, r_ijA score for item j on behalf of user i;

in the populated user-rating matrix R_m×nOn the basis, the similarity between users is calculated by using Pearson correlation, and a user-user similarity matrix S is constructed:

s_uvrepresents the degree of similarity between user u and user v, and s_uv∈[0,1]；

Calculating the trust degree, and then calculating a weighted value combining the similarity and the trust:

h(i,j)＝θ×trust(i,j)+(1-θ)×sim(i,j)；

theta represents a weighting parameter, trust (i, j) represents the direct trust of users i and j, sim (i, j) represents similarity, and h (i, j) represents a weighted value of the two;

and (3) constructing a user relationship matrix H by using the weighted values:

3. the information recommendation method based on the social network as claimed in claim 2, wherein the direct trust is a combination of the interactive trust, the proportion of common friends among users, and the evaluation ability of the users, and the calculation formula is as follows:

trust (i, j) represents the direct trust of users i and j, fri (i, j) represents the proportion of common friends between the users, fre represents the evaluation capability of the users, and comm (i, j) represents the interactive trust between the users;

the calculation formula of the interaction trust degree is as follows:

wherein, w_i,jRepresenting the number of messages sent by user i to user j, Σ w_i(out)The total number of messages sent to surrounding users on behalf of the user i;

the calculation formula of the proportion of common friends among users and the evaluation capability of the users is as follows:

where fri (i, j) represents the proportion of common friends among users, fre represents the evaluation ability of users, and n_iAnd n_jRepresenting the numbers of friends of users i and j, n_i∩n_jNumber of common friends on their behalf; k is a radical of_iRepresenting the number of evaluations of the recommended user on the i-type merchandise, Σ k_iRepresenting the number of evaluations of all the commodity categories by the user.

4. The method of claim 1, wherein the community discovery method in step2 comprises the following steps:

step 1: calculating the degree of each user in the network (the number of edges associated with the vertex), and selecting the user i with the highest degree from the degrees as an initial community C_iAnd initializing the modularity Q to be 0;

step 2: find all and community C_iConnected users and put them into the adjacent user set N;

step 3: calculating each user j in the user set N to the community C_iAnd adding the user having the largest contribution degree to the community C_iPerforming the following steps;

step 4: compute community C_iIf Q' is greater than Q, adding user j to community C_iSuccessfully marking the user j, updating the modularity Q ═ Q', and returning to Step2 to continue execution; otherwise, go to Step 5;

step 5: the modularity Q has reached a maximum value, i.e. the current community C_iAchieving the optimal result of the division;

step 6: if there are no unmarked users and all communities in the network have been detected, the process ends; otherwise, selecting the user with the maximum degree from the users without marks as a new initial community C_iReturning to step2 to continue execution.

5. The information recommendation method based on the social network as claimed in claim 4, wherein the contribution q of the user to the community is calculated as follows:

L_in: representing the number of connected edges in the community in the unauthorized network; representing the total weight of all edges in the community in the weighted network;

L_out: representing the number of external connection edges connected with the community in the unauthorized network; representing outside connections to a community in a privileged networkThe sum of the weights of all edges is calculated;

the larger the contribution q of the user to the community is, the more closely the user and the community are connected.

6. The information recommendation method based on social network as claimed in claim 4, wherein the modularity Q is expressed as follows:

if the network is an unweighted network, the expression of the modularity Q is as follows:

wherein m is the total number of edges of the network; k is a radical of_iAnd k_jRespectively representing the number of connecting edges of the user i and the user j; a. the_ijRepresenting a network adjacency matrix, A when a user i and a user j of a random network connection are connected_ijWhen user i and user j are not connected, a is 1_ij＝0；δ(C_i,C_j) For the kronecker function, if user i and user j belong to the same community, δ (C)_i,C_j) 1, if not in the same community, δ (C)_i,C_j)＝0；

In a weighted network, the modularity Q may be defined as follows:

wherein W represents the sum of the weights of all edges in the network, W_iAnd w_jRepresenting the sum of the weights of the edges connected to user i and user j, respectively.

7. The information recommendation method based on the social network as claimed in claim 1, wherein the scoring recommendation method in step3 is as follows: predicting the scoring of the candidate items by the target user according to the scoring information of the candidate items by the N nearest neighbors of the target user, selecting the first items with the highest predicted scores as recommendation results and actively pushing the recommendation results to the target user, namely generating top-N information resource recommendation:

wherein, P_u,iRepresents the predicted rating of item i by user u,andrespectively representing the average scores of the user u and the neighbor user v on the project; r_v,iRepresents the rating of the user v on the item i, trust (u, v) represents the trust degree of the user u on the neighbor user v, and N represents the neighbor user candidate set of the user u.

8. The method of claim 7, wherein the direct trust (u, v) is also the transitive trust (Trust)_L(A,B)Instead, it is calculated as follows:

trust_L(A,B)＝trust(A,X₁)×trust(X₁,X₂)×…×trust(X_n,B)；

wherein, X_iRepresenting users between users A and B on path L, L (A, B) representing an existing trust path between user A and user B, if there are multiple trust paths L (L) between user A and user B in the trust network₁,L₂,…,L_n) (n is more than or equal to 2), selecting the shortest path in the path L, if k shortest paths exist, calculating the formula as follows:

then, the calculation is continued according to the sequence of the path L from the short to the long, if the calculation result of a certain path jCalculate path j as one of the shortest paths and recalculate trust_L(A,B)And when all paths are calculated, a final indirect trust result is obtained.